Libraries

library(tidyverse)
library(mlbench)
library(ggfortify)
library(GGally)
library(scagnostics)
library(mlr) 

Dataset

Pima Indians Diabetes dataset from mlbench package.

data(PimaIndiansDiabetes)
PimaIndiansDiabetes %>%
  head()
##   pregnant glucose pressure triceps insulin mass pedigree age diabetes
## 1        6     148       72      35       0 33.6    0.627  50      pos
## 2        1      85       66      29       0 26.6    0.351  31      neg
## 3        8     183       64       0       0 23.3    0.672  32      pos
## 4        1      89       66      23      94 28.1    0.167  21      neg
## 5        0     137       40      35     168 43.1    2.288  33      pos
## 6        5     116       74       0       0 25.6    0.201  30      neg

Colors

# The palette with grey:
cbp1 <- c("#999999", "#E69F00", "#56B4E9", "#009E73",
          "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
ggplot <- function(...) ggplot2::ggplot(...) + 
  scale_color_manual(values = cbp1) +
  scale_fill_manual(values = cbp1) + # note: needs to be overridden when using continuous color scales
  theme_bw()

Exploratory Data Analysis

Exploratory Data Analysis (EDA) is the backbone of data analysis, including those that result in a machine learning model. EDA helps us to understand the data we are working with and put it into context, so that we are able to ask the right questions (or to put our questions into the right frame). It helps us take appropriate measures for cleaning, normalization/transformation, dealing with missing values, feature preparation and engineering, etc. Particularly if our machine learning model is trained on a limited dataset (but not only then!), appropriate data preparation can vastly improve the machine learning process: models will often train faster and achieve higher accuracy.

An essential part of EDA is data visualization.

Typically, we want to start by exploring potential sources of errors in our data, like

  • wrong/useless data types (sometimes data types are automatically set in a way that is not useful for our analysis, like factors versus strings, or wrong/strange entries in an otherwise numeric column will make it categorical)
  • missing values (a collection of ways to visualize missingness can be found here),
  • outliers (for example by plotting a box-plot of continuous variables)

Depending on the number of features/variables we have, it makes sense to look at them all individually and in correlation with each other. Depending on whether we have a categorical or continuous variable, we might be interested in properties that are shown by

  • histograms (frequency distribution of binned continuous variables),
  • density distribution (normalized distribution of continuous variables) or
  • bar-plots (shows counts of categorical variables).

If our target variable is categorical, we will want to look at potential imbalances between the classes. Class imbalance will strongly affect the machine learning modeling process and will require us to consider up-/downsampling or similar techniques before we train a model.

Correlation analysis can show us, for example

  • how our target/dependent variable correlates with the remaining features (often, just by looking at the correlation, we can identify one ore more feature that will have a strong impact on predicting the target because they are strongly correlated) or
  • whether some of the independent variables/features correlate with each other (multicolinearity; we might want to consider removing strongly correlated features, so that they won’t contribute the “same” information multiple times to the model and thus lead to overfitting).

Additional methods can be used to visualize groups of related features. These methods are often especially useful if we have a large dataset with a large feature set (highly dimensional data). Some of these methods for visualizing groups of related features and/or for comparing multiple variables and visualizing their relationships are:

  • Dimensionality reduction:
    • Principal Component Analysis (PCA, linear, shows as much variation in data as possible)
    • Multidimensional scaling (MDS, non-linear)
    • Sammon mapping (non-linear)
    • T-Distributed Stochastic Neighbor Embedding (t-SNE, non-linear)
    • Uniform Manifold Approximation and Projection (UMAP, non-linear, faster than T-SNE, often captures global variation better than T-SNE and PCA)
    • Isometric Feature Mapping Ordination (Isomap)
  • Parallel coordinate plots
  • scagnostics
# in our dataset,
# continuous variables are
PimaIndiansDiabetes %>%
  dplyr::select(where(is.numeric)) %>%
  head()
##   pregnant glucose pressure triceps insulin mass pedigree age
## 1        6     148       72      35       0 33.6    0.627  50
## 2        1      85       66      29       0 26.6    0.351  31
## 3        8     183       64       0       0 23.3    0.672  32
## 4        1      89       66      23      94 28.1    0.167  21
## 5        0     137       40      35     168 43.1    2.288  33
## 6        5     116       74       0       0 25.6    0.201  30
# 'diabetes' is the only categorical variable is also our target or dependent variable
PimaIndiansDiabetes %>%
  dplyr::select(!where(is.numeric)) %>%
  head()
##   diabetes
## 1      pos
## 2      neg
## 3      pos
## 4      neg
## 5      pos
## 6      neg
# bar plot of target
PimaIndiansDiabetes %>%
  ggplot(aes(x = diabetes, fill = diabetes)) +
    geom_bar(alpha = 0.8) +
    theme(legend.position = "none") +
    labs(x = "Diabetes outcome", 
         y = "count",
        title = "Barplot of categorical features", 
        caption = "Source: Pima Indians Diabetes Database")

# boxplot of continuous features
PimaIndiansDiabetes %>%
  gather("key", "value", pregnant:age) %>%
  ggplot(aes(x = value, fill = diabetes)) +
    facet_wrap(vars(key), ncol = 3, scales = "free") +
    geom_boxplot(alpha = 0.8) +
    theme(axis.text.y = element_blank(),
          axis.ticks.y = element_blank())

# histogram of features
PimaIndiansDiabetes %>%
  gather("key", "value", pregnant:age) %>%
  ggplot(aes(x = value, fill = diabetes)) +
    facet_wrap(vars(key), ncol = 3, scales = "free") +
    geom_histogram(alpha = 0.8) +
    labs(x = "value of feature in facet", 
         y = "count",
         fill = "Diabetes",
        title = "Histogram of features", 
        caption = "Source: Pima Indians Diabetes Database")

# density plot of of features
PimaIndiansDiabetes %>%
  gather("key", "value", pregnant:age) %>%
  ggplot(aes(x = value, fill = diabetes)) +
    facet_wrap(vars(key), ncol = 3, scales = "free") +
    geom_density(alpha = 0.8) +
    labs(x = "value of feature in facet", 
         y = "density",
         fill = "Diabetes",
        title = "Density of continuous features", 
        caption = "Source: Pima Indians Diabetes Database")

# correlation plot of features
mat <- PimaIndiansDiabetes %>%
  dplyr::select(where(is.numeric))

cormat <- round(cor(mat), 2)

cormat <- cormat %>%
  as_data_frame() %>%
  mutate(x = colnames(mat)) %>%
  gather(key = "y", value = "value", pregnant:age)

cormat %>%
    remove_missing() %>%
    arrange(x, y) %>%
    ggplot(aes(x = x, y = y, fill = value)) + 
    geom_tile() +
    scale_fill_gradient2(low = "blue", high = "red", mid = "white", 
     midpoint = 0, limit = c(-1,1), space = "Lab", 
     name = "Pearson\nCorrelation") +
    theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust = 1)) +
    coord_fixed() +
    labs(x = "feature", 
         y = "feature",
        title = "Correlation between features", 
        caption = "Source: Pima Indians Diabetes Database")

# scatterplot matrix
ggpairs(PimaIndiansDiabetes, 
        columns = c(1:8),
        alpha = 0.7) +
    labs(x = "feature", 
         y = "feature",
        title = "Scatterplot matrix", 
        caption = "Source: Pima Indians Diabetes Database")

# PCA
prep <- PimaIndiansDiabetes %>%
  dplyr::select(where(is.numeric))

pca <- prep %>%
  prcomp(scale. = TRUE)

autoplot(pca, 
                data = PimaIndiansDiabetes, 
                colour = 'diabetes',
                shape = 'diabetes',
                loadings = TRUE, 
                loadings.colour = 'blue',
                loadings.label = TRUE, 
                loadings.label.size = 3) +
      scale_color_manual(values = cbp1) +
  scale_fill_manual(values = cbp1) +
  theme_bw() +
    labs(title = "Principal Component Analysis (PCA)", 
        caption = "Source: Pima Indians Diabetes Database")

# MDS
d <- dist(prep) # euclidean distances between the rows
fit <- cmdscale(d,eig=TRUE, k=2) # k is the number of dim
fit$points %>%
  head()
##        [,1]       [,2]
## 1 -75.71465 -35.950783
## 2 -82.35827  28.908213
## 3 -74.63064 -67.906496
## 4  11.07742  34.898486
## 5  89.74379  -2.746937
## 6 -80.97792  -3.946887
# Sammon mapping
library(MASS)
sam <- sammon(dist(prep))
## Initial stress        : 0.03033
## stress after   0 iters: 0.03033
sam$points %>%
  head()
##        [,1]       [,2]
## 1 -75.71465 -35.950783
## 2 -82.35827  28.908213
## 3 -74.63064 -67.906496
## 4  11.07742  34.898486
## 5  89.74379  -2.746937
## 6 -80.97792  -3.946887
# parallel coordinate plots
ggparcoord(data = PimaIndiansDiabetes, 
           columns = c(1:8), 
           groupColumn = 9,
           scale = "robust",
           order = "skewness",
           alpha = 0.7)

# scagnostics
scagnostics_dataset <- scagnostics(PimaIndiansDiabetes)

# scagnostics grid
scagnostics_grid_dataset <- scagnosticsGrid(scagnostics_dataset)

# outliers
scagnostics_o_dataset <- scagnosticsOutliers(scagnostics_dataset)
scagnostics_o_dataset[scagnostics_o_dataset]
## pregnant * age 
##           TRUE
outlier <- scagnostics_grid_dataset[scagnostics_o_dataset,]

# scagnostics exemplars
scagnostics_ex_dataset <- scagnosticsExemplars(scagnostics_dataset)
scagnostics_ex_dataset[scagnostics_ex_dataset]
## pregnant * triceps         mass * age triceps * diabetes 
##               TRUE               TRUE               TRUE
exemplars <- scagnostics_grid_dataset[scagnostics_ex_dataset,]

Training a machine learning model

(using mlr package)

  • create training and test set
set.seed(1000) 

train_index <- sample(1:nrow(PimaIndiansDiabetes), 0.8 * nrow(PimaIndiansDiabetes)) 
test_index <- setdiff(1:nrow(PimaIndiansDiabetes), train_index) 

train <- PimaIndiansDiabetes[train_index,] 
test <- PimaIndiansDiabetes[test_index,]

list( train = summary(train), test = summary(test) )
## $train
##     pregnant         glucose         pressure         triceps     
##  Min.   : 0.000   Min.   :  0.0   Min.   :  0.00   Min.   : 0.00  
##  1st Qu.: 1.000   1st Qu.:100.0   1st Qu.: 64.00   1st Qu.: 0.00  
##  Median : 3.000   Median :119.0   Median : 72.00   Median :23.00  
##  Mean   : 3.894   Mean   :123.1   Mean   : 68.89   Mean   :20.66  
##  3rd Qu.: 6.000   3rd Qu.:143.0   3rd Qu.: 80.00   3rd Qu.:32.75  
##  Max.   :17.000   Max.   :199.0   Max.   :114.00   Max.   :99.00  
##     insulin            mass          pedigree           age        diabetes 
##  Min.   :  0.00   Min.   : 0.00   Min.   :0.0780   Min.   :21.00   neg:386  
##  1st Qu.:  0.00   1st Qu.:27.10   1st Qu.:0.2442   1st Qu.:24.00   pos:228  
##  Median : 36.50   Median :32.00   Median :0.3780   Median :29.00            
##  Mean   : 81.65   Mean   :31.92   Mean   :0.4742   Mean   :33.42            
##  3rd Qu.:131.50   3rd Qu.:36.38   3rd Qu.:0.6355   3rd Qu.:41.00            
##  Max.   :846.00   Max.   :59.40   Max.   :2.4200   Max.   :81.00            
## 
## $test
##     pregnant         glucose         pressure         triceps     
##  Min.   : 0.000   Min.   :  0.0   Min.   :  0.00   Min.   : 0.00  
##  1st Qu.: 1.000   1st Qu.: 93.0   1st Qu.: 62.00   1st Qu.: 0.00  
##  Median : 2.000   Median :108.0   Median : 72.00   Median :23.00  
##  Mean   : 3.649   Mean   :112.3   Mean   : 69.96   Mean   :20.03  
##  3rd Qu.: 6.000   3rd Qu.:133.8   3rd Qu.: 79.50   3rd Qu.:32.00  
##  Max.   :14.000   Max.   :197.0   Max.   :122.00   Max.   :56.00  
##     insulin           mass          pedigree           age        diabetes 
##  Min.   :  0.0   Min.   : 0.00   Min.   :0.0850   Min.   :21.00   neg:114  
##  1st Qu.:  0.0   1st Qu.:27.80   1st Qu.:0.2395   1st Qu.:23.25   pos: 40  
##  Median : 20.5   Median :32.40   Median :0.3380   Median :29.00            
##  Mean   : 72.4   Mean   :32.29   Mean   :0.4627   Mean   :32.54            
##  3rd Qu.:100.0   3rd Qu.:36.88   3rd Qu.:0.6008   3rd Qu.:39.75            
##  Max.   :744.0   Max.   :67.10   Max.   :2.3290   Max.   :67.00
  • create classification task and learner
listLearners()
##                              class
## 1                      classif.ada
## 2               classif.adaboostm1
## 3              classif.bartMachine
## 4                 classif.binomial
## 5                 classif.boosting
## 6                      classif.bst
## 7                      classif.C50
## 8                  classif.cforest
## 9               classif.clusterSVM
## 10                   classif.ctree
## 11                classif.cvglmnet
## 12                  classif.dbnDNN
## 13                   classif.dcSVM
## 14                   classif.earth
## 15                  classif.evtree
## 16              classif.extraTrees
## 17              classif.fdausc.glm
## 18           classif.fdausc.kernel
## 19              classif.fdausc.knn
## 20               classif.fdausc.np
## 21                 classif.FDboost
## 22             classif.featureless
## 23                    classif.fgam
## 24                     classif.fnn
## 25                classif.gamboost
## 26                classif.gaterSVM
## 27                 classif.gausspr
## 28                     classif.gbm
## 29                   classif.geoDA
## 30                classif.glmboost
## 31                  classif.glmnet
## 32        classif.h2o.deeplearning
## 33                 classif.h2o.gbm
## 34                 classif.h2o.glm
## 35        classif.h2o.randomForest
## 36                     classif.IBk
## 37                     classif.J48
## 38                    classif.JRip
## 39                    classif.kknn
## 40                     classif.knn
## 41                    classif.ksvm
## 42                     classif.lda
## 43        classif.LiblineaRL1L2SVC
## 44       classif.LiblineaRL1LogReg
## 45        classif.LiblineaRL2L1SVC
## 46       classif.LiblineaRL2LogReg
## 47          classif.LiblineaRL2SVC
## 48  classif.LiblineaRMultiClassSVC
## 49                   classif.linDA
## 50                  classif.logreg
## 51                   classif.lssvm
## 52                    classif.lvq1
## 53                     classif.mda
## 54                     classif.mlp
## 55                classif.multinom
## 56              classif.naiveBayes
## 57               classif.neuralnet
## 58                    classif.nnet
## 59                 classif.nnTrain
## 60             classif.nodeHarvest
## 61                    classif.OneR
## 62                    classif.pamr
## 63                    classif.PART
## 64               classif.penalized
## 65                     classif.plr
## 66              classif.plsdaCaret
## 67                  classif.probit
## 68                     classif.qda
## 69                   classif.quaDA
## 70            classif.randomForest
## 71         classif.randomForestSRC
## 72                  classif.ranger
## 73                     classif.rda
## 74                  classif.rFerns
## 75                    classif.rknn
## 76          classif.rotationForest
## 77                   classif.rpart
## 78                     classif.RRF
## 79                   classif.rrlda
## 80                  classif.saeDNN
## 81                     classif.sda
## 82               classif.sparseLDA
## 83                     classif.svm
## 84                 classif.xgboost
## 85                  cluster.cmeans
## 86                  cluster.Cobweb
## 87                  cluster.dbscan
## 88                      cluster.EM
## 89           cluster.FarthestFirst
## 90                 cluster.kkmeans
## 91                  cluster.kmeans
## 92         cluster.MiniBatchKmeans
## 93            cluster.SimpleKMeans
## 94                  cluster.XMeans
## 95              multilabel.cforest
## 96      multilabel.randomForestSRC
## 97               multilabel.rFerns
## 98                regr.bartMachine
## 99                      regr.bcart
## 100                       regr.bgp
## 101                    regr.bgpllm
## 102                       regr.blm
## 103                      regr.brnn
## 104                       regr.bst
## 105                      regr.btgp
## 106                   regr.btgpllm
## 107                      regr.btlm
## 108                   regr.cforest
## 109                       regr.crs
## 110                     regr.ctree
## 111                    regr.cubist
## 112                  regr.cvglmnet
## 113                     regr.earth
## 114                    regr.evtree
## 115                regr.extraTrees
## 116                   regr.FDboost
## 117               regr.featureless
## 118                      regr.fgam
## 119                       regr.fnn
## 120                      regr.frbs
## 121                  regr.gamboost
## 122                   regr.gausspr
## 123                       regr.gbm
## 124                       regr.glm
## 125                  regr.glmboost
## 126                    regr.glmnet
## 127                     regr.GPfit
## 128          regr.h2o.deeplearning
## 129                   regr.h2o.gbm
## 130                   regr.h2o.glm
## 131          regr.h2o.randomForest
## 132                       regr.IBk
## 133                      regr.kknn
## 134                        regr.km
## 135                      regr.ksvm
## 136                      regr.laGP
## 137          regr.LiblineaRL2L1SVR
## 138          regr.LiblineaRL2L2SVR
## 139                        regr.lm
## 140                      regr.mars
## 141                       regr.mob
## 142                      regr.nnet
## 143               regr.nodeHarvest
## 144                       regr.pcr
## 145                 regr.penalized
## 146                      regr.plsr
## 147              regr.randomForest
## 148           regr.randomForestSRC
## 149                    regr.ranger
## 150                      regr.rknn
## 151                     regr.rpart
## 152                       regr.RRF
## 153                       regr.rsm
## 154                       regr.rvm
## 155                       regr.svm
## 156                   regr.xgboost
## 157                   surv.cforest
## 158                     surv.coxph
## 159                  surv.cvglmnet
## 160                  surv.gamboost
## 161                       surv.gbm
## 162                  surv.glmboost
## 163                    surv.glmnet
## 164           surv.randomForestSRC
## 165                    surv.ranger
## 166                     surv.rpart
##                                                                                                          name
## 1                                                                                                ada Boosting
## 2                                                                                             ada Boosting M1
## 3                                                                          Bayesian Additive Regression Trees
## 4                                                                                         Binomial Regression
## 5                                                                                             Adabag Boosting
## 6                                                                                           Gradient Boosting
## 7                                                                                                         C50
## 8                                                          Random forest based on conditional inference trees
## 9                                                                           Clustered Support Vector Machines
## 10                                                                                Conditional Inference Trees
## 11                                       GLM with Lasso or Elasticnet Regularization (Cross Validated Lambda)
## 12                                                        Deep neural network with weights initialized by DBN
## 13                                                                    Divided-Conquer Support Vector Machines
## 14                                                                             Flexible Discriminant Analysis
## 15                                                            Evolutionary learning of globally optimal trees
## 16                                                                                 Extremely Randomized Trees
## 17                                                            Generalized Linear Models classification on FDA
## 18                                                                               Kernel classification on FDA
## 19                                                                                                 fdausc.knn
## 20                                                                        Nonparametric classification on FDA
## 21                                                            Functional linear array classification boosting
## 22                                                                                     Featureless classifier
## 23                                                                          functional general additive model
## 24                                                                                   Fast k-Nearest Neighbour
## 25                                                                   Gradient boosting with smooth components
## 26                                                         Mixture of SVMs with Neural Network Gater Function
## 27                                                                                         Gaussian Processes
## 28                                                                                  Gradient Boosting Machine
## 29                                                                 Geometric Predictive Discriminant Analysis
## 30                                                                                          Boosting for GLMs
## 31                                                                GLM with Lasso or Elasticnet Regularization
## 32                                                                                           h2o.deeplearning
## 33                                                                                                    h2o.gbm
## 34                                                                                                    h2o.glm
## 35                                                                                           h2o.randomForest
## 36                                                                                       k-Nearest Neighbours
## 37                                                                                         J48 Decision Trees
## 38                                                                                 Propositional Rule Learner
## 39                                                                                         k-Nearest Neighbor
## 40                                                                                         k-Nearest Neighbor
## 41                                                                                    Support Vector Machines
## 42                                                                               Linear Discriminant Analysis
## 43                                                       L1-Regularized L2-Loss Support Vector Classification
## 44                                                                         L1-Regularized Logistic Regression
## 45                                                       L2-Regularized L1-Loss Support Vector Classification
## 46                                                                         L2-Regularized Logistic Regression
## 47                                                       L2-Regularized L2-Loss Support Vector Classification
## 48                                                        Support Vector Classification by Crammer and Singer
## 49                                                                               Linear Discriminant Analysis
## 50                                                                                        Logistic Regression
## 51                                                                       Least Squares Support Vector Machine
## 52                                                                               Learning Vector Quantization
## 53                                                                              Mixture Discriminant Analysis
## 54                                                                                     Multi-Layer Perceptron
## 55                                                                                     Multinomial Regression
## 56                                                                                                Naive Bayes
## 57                                                                              Neural Network from neuralnet
## 58                                                                                             Neural Network
## 59                                                                 Training Neural Network by Backpropagation
## 60                                                                                               Node Harvest
## 61                                                                                             1-R Classifier
## 62                                                                                  Nearest shrunken centroid
## 63                                                                                        PART Decision Lists
## 64                                                                              Penalized Logistic Regression
## 65                                                                      Logistic Regression with a L2 Penalty
## 66                                                          Partial Least Squares (PLS) Discriminant Analysis
## 67                                                                                          Probit Regression
## 68                                                                            Quadratic Discriminant Analysis
## 69                                                                            Quadratic Discriminant Analysis
## 70                                                                                              Random Forest
## 71                                                                                              Random Forest
## 72                                                                                             Random Forests
## 73                                                                          Regularized Discriminant Analysis
## 74                                                                                               Random ferns
## 75                                                                                 Random k-Nearest-Neighbors
## 76                                                                                            Rotation Forest
## 77                                                                                              Decision Tree
## 78                                                                                 Regularized Random Forests
## 79                                                            Robust Regularized Linear Discriminant Analysis
## 80                                        Deep neural network with weights initialized by Stacked AutoEncoder
## 81                                                                            Shrinkage Discriminant Analysis
## 82                                                                               Sparse Discriminant Analysis
## 83                                                                           Support Vector Machines (libsvm)
## 84                                                                                  eXtreme Gradient Boosting
## 85                                                                                   Fuzzy C-Means Clustering
## 86                                                                                Cobweb Clustering Algorithm
## 87                                                                                          DBScan Clustering
## 88                                                                        Expectation-Maximization Clustering
## 89                                                                         FarthestFirst Clustering Algorithm
## 90                                                                                             Kernel K-Means
## 91                                                                                                    K-Means
## 92                                                                                            MiniBatchKmeans
## 93                                                                                         K-Means Clustering
## 94                                                         XMeans (k-means with automatic determination of k)
## 95                                                         Random forest based on conditional inference trees
## 96                                                                                              Random Forest
## 97                                                                                               Random ferns
## 98                                                                         Bayesian Additive Regression Trees
## 99                                                                                              Bayesian CART
## 100                                                                                 Bayesian Gaussian Process
## 101                                         Bayesian Gaussian Process with jumps to the Limiting Linear Model
## 102                                                                                     Bayesian Linear Model
## 103                                                  Bayesian regularization for feed-forward neural networks
## 104                                                                                         Gradient Boosting
## 105                                                                           Bayesian Treed Gaussian Process
## 106                                   Bayesian Treed Gaussian Process with jumps to the Limiting Linear Model
## 107                                                                               Bayesian Treed Linear Model
## 108                                                        Random Forest Based on Conditional Inference Trees
## 109                                                                                        Regression Splines
## 110                                                                               Conditional Inference Trees
## 111                                                                                                    Cubist
## 112                                      GLM with Lasso or Elasticnet Regularization (Cross Validated Lambda)
## 113                                                                  Multivariate Adaptive Regression Splines
## 114                                                           Evolutionary learning of globally optimal trees
## 115                                                                                Extremely Randomized Trees
## 116                                                               Functional linear array regression boosting
## 117                                                                                    Featureless regression
## 118                                                                         functional general additive model
## 119                                                                                   Fast k-Nearest Neighbor
## 120                                                                                  Fuzzy Rule-based Systems
## 121                                                                  Gradient Boosting with Smooth Components
## 122                                                                                        Gaussian Processes
## 123                                                                                 Gradient Boosting Machine
## 124                                                                             Generalized Linear Regression
## 125                                                                                         Boosting for GLMs
## 126                                                               GLM with Lasso or Elasticnet Regularization
## 127                                                                                          Gaussian Process
## 128                                                                                          h2o.deeplearning
## 129                                                                                                   h2o.gbm
## 130                                                                                                   h2o.glm
## 131                                                                                          h2o.randomForest
## 132                                                                                      K-Nearest Neighbours
## 133                                                                             K-Nearest-Neighbor regression
## 134                                                                                                   Kriging
## 135                                                                                   Support Vector Machines
## 136                                                                        Local Approximate Gaussian Process
## 137                                                          L2-Regularized L1-Loss Support Vector Regression
## 138                                                          L2-Regularized L2-Loss Support Vector Regression
## 139                                                                                  Simple Linear Regression
## 140                                                                  Multivariate Adaptive Regression Splines
## 141 Model-based Recursive Partitioning  Yielding a Tree with Fitted Models Associated with each Terminal Node
## 142                                                                                            Neural Network
## 143                                                                                              Node Harvest
## 144                                                                            Principal Component Regression
## 145                                                                                      Penalized Regression
## 146                                                                          Partial Least Squares Regression
## 147                                                                                             Random Forest
## 148                                                                                             Random Forest
## 149                                                                                            Random Forests
## 150                                                                                Random k-Nearest-Neighbors
## 151                                                                                             Decision Tree
## 152                                                                                Regularized Random Forests
## 153                                                                               Response Surface Regression
## 154                                                                                  Relevance Vector Machine
## 155                                                                          Support Vector Machines (libsvm)
## 156                                                                                 eXtreme Gradient Boosting
## 157                                                        Random Forest based on Conditional Inference Trees
## 158                                                                             Cox Proportional Hazard Model
## 159                                                          GLM with Regularization (Cross Validated Lambda)
## 160                                                                  Gradient boosting with smooth components
## 161                                                                                 Gradient Boosting Machine
## 162                                                        Gradient Boosting with Componentwise Linear Models
## 163                                                                                   GLM with Regularization
## 164                                                                                             Random Forest
## 165                                                                                            Random Forests
## 166                                                                                             Survival Tree
##              short.name                   package
## 1                   ada                 ada,rpart
## 2            adaboostm1                     RWeka
## 3           bartmachine               bartMachine
## 4              binomial                     stats
## 5                adabag              adabag,rpart
## 6                   bst                 bst,rpart
## 7                   C50                       C50
## 8               cforest                     party
## 9            clusterSVM        SwarmSVM,LiblineaR
## 10                ctree                     party
## 11             cvglmnet                    glmnet
## 12              dbn.dnn                   deepnet
## 13                dcSVM            SwarmSVM,e1071
## 14                  fda               earth,stats
## 15               evtree                    evtree
## 16           extraTrees                extraTrees
## 17           fdausc.glm                   fda.usc
## 18        fdausc.kernel                   fda.usc
## 19           fdausc.knn                   fda.usc
## 20            fdausc.np                   fda.usc
## 21              FDboost            FDboost,mboost
## 22          featureless                       mlr
## 23                 FGAM                    refund
## 24                  fnn                       FNN
## 25             gamboost                    mboost
## 26             gaterSVM                  SwarmSVM
## 27              gausspr                   kernlab
## 28                  gbm                       gbm
## 29                geoda               DiscriMiner
## 30             glmboost                    mboost
## 31               glmnet                    glmnet
## 32               h2o.dl                       h2o
## 33              h2o.gbm                       h2o
## 34              h2o.glm                       h2o
## 35               h2o.rf                       h2o
## 36                  ibk                     RWeka
## 37                  j48                     RWeka
## 38                 jrip                     RWeka
## 39                 kknn                      kknn
## 40                  knn                     class
## 41                 ksvm                   kernlab
## 42                  lda                      MASS
## 43        liblinl1l2svc                 LiblineaR
## 44       liblinl1logreg                 LiblineaR
## 45        liblinl2l1svc                 LiblineaR
## 46       liblinl2logreg                 LiblineaR
## 47          liblinl2svc                 LiblineaR
## 48  liblinmulticlasssvc                 LiblineaR
## 49                linda               DiscriMiner
## 50               logreg                     stats
## 51                lssvm                   kernlab
## 52                 lvq1                     class
## 53                  mda                       mda
## 54                  mlp                     RSNNS
## 55             multinom                      nnet
## 56               nbayes                     e1071
## 57            neuralnet                 neuralnet
## 58                 nnet                      nnet
## 59             nn.train                   deepnet
## 60          nodeHarvest               nodeHarvest
## 61                 oner                     RWeka
## 62                 pamr                      pamr
## 63                 part                     RWeka
## 64            penalized                 penalized
## 65                  plr                   stepPlr
## 66           plsdacaret                 caret,pls
## 67               probit                     stats
## 68                  qda                      MASS
## 69                quada               DiscriMiner
## 70                   rf              randomForest
## 71                rfsrc           randomForestSRC
## 72               ranger                    ranger
## 73                  rda                      klaR
## 74               rFerns                    rFerns
## 75                 rknn                      rknn
## 76       rotationForest            rotationForest
## 77                rpart                     rpart
## 78                  RRF                       RRF
## 79                rrlda                     rrlda
## 80              sae.dnn                   deepnet
## 81                  sda                       sda
## 82            sparseLDA sparseLDA,MASS,elasticnet
## 83                  svm                     e1071
## 84              xgboost                   xgboost
## 85               cmeans                e1071,clue
## 86               cobweb                     RWeka
## 87               dbscan                       fpc
## 88                   em                     RWeka
## 89        farthestfirst                     RWeka
## 90              kkmeans                   kernlab
## 91               kmeans                stats,clue
## 92         MBatchKmeans                  ClusterR
## 93         simplekmeans                     RWeka
## 94               xmeans                     RWeka
## 95              cforest                     party
## 96                rfsrc           randomForestSRC
## 97               rFerns                    rFerns
## 98          bartmachine               bartMachine
## 99                bcart                       tgp
## 100                 bgp                       tgp
## 101              bgpllm                       tgp
## 102                 blm                       tgp
## 103                brnn                      brnn
## 104                 bst                 bst,rpart
## 105                btgp                       tgp
## 106             btgpllm                       tgp
## 107                btlm                       tgp
## 108             cforest                     party
## 109                 crs                       crs
## 110               ctree                     party
## 111              cubist                    Cubist
## 112            cvglmnet                    glmnet
## 113               earth                     earth
## 114              evtree                    evtree
## 115          extraTrees                extraTrees
## 116             FDboost            FDboost,mboost
## 117         featureless                       mlr
## 118                FGAM                    refund
## 119                 fnn                       FNN
## 120                frbs                      frbs
## 121            gamboost                    mboost
## 122             gausspr                   kernlab
## 123                 gbm                       gbm
## 124                 glm                     stats
## 125            glmboost                    mboost
## 126              glmnet                    glmnet
## 127               GPfit                     GPfit
## 128              h2o.dl                       h2o
## 129             h2o.gbm                       h2o
## 130             h2o.glm                       h2o
## 131              h2o.rf                       h2o
## 132                 ibk                     RWeka
## 133                kknn                      kknn
## 134                  km               DiceKriging
## 135                ksvm                   kernlab
## 136                laGP                      laGP
## 137       liblinl2l1svr                 LiblineaR
## 138       liblinl2l2svr                 LiblineaR
## 139                  lm                     stats
## 140                mars                       mda
## 141                 mob          party,modeltools
## 142                nnet                      nnet
## 143         nodeHarvest               nodeHarvest
## 144                 pcr                       pls
## 145           penalized                 penalized
## 146                plsr                       pls
## 147                  rf              randomForest
## 148               rfsrc           randomForestSRC
## 149              ranger                    ranger
## 150                rknn                      rknn
## 151               rpart                     rpart
## 152                 RRF                       RRF
## 153                 rsm                       rsm
## 154                 rvm                   kernlab
## 155                 svm                     e1071
## 156             xgboost                   xgboost
## 157                 crf            party,survival
## 158               coxph                  survival
## 159            cvglmnet                    glmnet
## 160            gamboost           survival,mboost
## 161                 gbm                       gbm
## 162            glmboost           survival,mboost
## 163              glmnet                    glmnet
## 164               rfsrc  survival,randomForestSRC
## 165              ranger                    ranger
## 166               rpart                     rpart
##                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               note
## 1                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 `xval` has been set to `0` by default for speed.
## 2                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      NAs are directly passed to WEKA with `na.action = na.pass`.
## 3                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              `use_missing_data` has been set to `TRUE` by default to allow missing data support.
## 4                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 Delegates to `glm` with freely choosable binomial link function via learner parameter `link`. We set 'model' to FALSE by default to save memory.
## 5                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 `xval` has been set to `0` by default for speed.
## 6                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                Renamed parameter `learner` to `Learner` due to nameclash with `setHyperPars`. Default changes: `Learner = "ls"`, `xval = 0`, and `maxdepth = 1`.
## 7                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 
## 8                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                See `?ctree_control` for possible breakage for nominal features with missingness.
## 9                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 `centers` set to `2` by default.
## 10                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               See `?ctree_control` for possible breakage for nominal features with missingness.
## 11                                                                                                                                                                  The family parameter is set to `binomial` for two-class problems and to `multinomial` otherwise. Factors automatically get converted to dummy columns, ordered factors to integer.\n      glmnet uses a global control object for its parameters. mlr resets all control parameters to their defaults\n      before setting the specified parameters and after training.\n      If you are setting glmnet.control parameters through glmnet.control,\n      you need to save and re-set them after running the glmnet learner.
## 12                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         `output` set to `"softmax"` by default.
## 13                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 14                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         This learner performs flexible discriminant analysis using the earth algorithm. na.action is set to na.fail and only this is supported.
## 15                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               `pmutatemajor`, `pmutateminor`, `pcrossover`, `psplit`, and `pprune`,\n      are scaled internally to sum to 100.
## 16                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 17                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       model$C[[1]] is set to quote(classif.glm)
## 18                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         Argument draw=FALSE is used as default.
## 19                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         Argument draw=FALSE is used as default.
## 20                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    Argument draw=FALSE is used as default. Additionally, mod$C[[1]] is set to quote(classif.np)
## 21                                                                                                                                                                                                                                                                                                                                                                                                      Uses only one base learner per functional or scalar covariate.\n      Uses the same hyperparameters for every baselearner.\n      Currently does not support interaction between scalar covariates.\n      Default for family has been set to 'Binomial', as 'Gaussian' is not applicable.
## 22                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 23                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 24                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 25                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            `family` has been set to `Binomial()` by default. For 'family' 'AUC' and 'AdaExp' probabilities cannot be predicted.
## 26                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            `m` set to `3` and `max.iter` set to `1` by default.
## 27                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  Kernel parameters have to be passed directly and not by using the `kpar` list in `gausspr`.\n    Note that `fit` has been set to `FALSE` by default for speed.
## 28                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       `keep.data` is set to FALSE to reduce memory requirements.\nParam 'n.cores' has been to set to '1' by default to suppress parallelization by the package.
## 29                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 30                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              `family` has been set to `Binomial` by default. For 'family' 'AUC' and 'AdaExp' probabilities cannot be predcited.
## 31  The family parameter is set to `binomial` for two-class problems and to `multinomial` otherwise.\n      Factors automatically get converted to dummy columns, ordered factors to integer.\n      Parameter `s` (value of the regularization parameter used for predictions) is set to `0.01` by default,\n      but needs to be tuned by the user.\n      glmnet uses a global control object for its parameters. mlr resets all control parameters to their defaults\n      before setting the specified parameters and after training.\n      If you are setting glmnet.control parameters through glmnet.control,\n      you need to save and re-set them after running the glmnet learner.
## 32                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         The default value of `missing_values_handling` is `"MeanImputation"`, so missing values are automatically mean-imputed.
## 33                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              'distribution' is set automatically to 'gaussian'.
## 34                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      `family` is always set to `"binomial"` to get a binary classifier. The default value of `missing_values_handling` is `"MeanImputation"`, so missing values are automatically mean-imputed.
## 35                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 36                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 37                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     NAs are directly passed to WEKA with `na.action = na.pass`.
## 38                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     NAs are directly passed to WEKA with `na.action = na.pass`.
## 39                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 40                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 41                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          Kernel parameters have to be passed directly and not by using the `kpar` list in `ksvm`. Note that `fit` has been set to `FALSE` by default for speed.
## 42                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           Learner parameter `predict.method` maps to `method` in `predict.lda`.
## 43                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 44                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 45                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 46                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              `type = 0` (the default) is primal and `type = 7` is dual problem.
## 47                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              `type = 2` (the default) is primal and `type = 1` is dual problem.
## 48                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 49                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     Set `validation = NULL` by default to disable internal test set validation.
## 50                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 Delegates to `glm` with `family = binomial(link = 'logit')`. We set 'model' to FALSE by default to save memory.
## 51                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          `fitted` has been set to `FALSE` by default for speed.
## 52                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 53                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 `keep.fitted` has been set to `FALSE` by default for speed and we use `start.method = "lvq"` for more robust behavior / less technical crashes.
## 54                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 55                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 56                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 57                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               `err.fct` has been set to `ce` and `linear.output` to FALSE to do classification.
## 58                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               `linout=TRUE` is hardcoded for regression. `size` has been set to `3` by default.
## 59                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                `output` set to `softmax` by default. `max.number.of.layers` can be set to control and tune the maximal number of layers specified via `hidden`.
## 60                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 61                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     NAs are directly passed to WEKA with `na.action = na.pass`.
## 62                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  Threshold for prediction (`threshold.predict`) has been set to `1` by default.
## 63                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     NAs are directly passed to WEKA with `na.action = na.pass`.
## 64                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       trace=FALSE was set by default to disable logging output.
## 65                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      AIC and BIC penalty types can be selected via the new parameter `cp.type`.
## 66                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 67                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                Delegates to `glm` with `family = binomial(link = 'probit')`. We set 'model' to FALSE by default to save memory.
## 68                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           Learner parameter `predict.method` maps to `method` in `predict.qda`.
## 69                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 70                                                                                                                                                                                                                                                                                                                                                                                                                                                        Note that the rf can freeze the R process if trained on a task with 1 feature which is constant. This can happen in feature forward selection, also due to resampling, and you need to remove such features with removeConstantFeatures.
## 71                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             `na.action` has been set to `"na.impute"` by default to allow missing data support.
## 72                                                                                                                                                                                                                                                                                                                                                                                                                                                           By default, internal parallelization is switched off (`num.threads = 1`), `verbose` output is disabled, `respect.unordered.factors` is set to `order` for all splitrules. If predict.type='prob' we set 'probability=TRUE' in ranger.
## 73                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  `estimate.error` has been set to `FALSE` by default for speed.
## 74                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 75                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                k restricted to < 99 as the code allocates arrays of static size
## 76                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 77                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                `xval` has been set to `0` by default for speed.
## 78                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 79                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 80                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         `output` set to `"softmax"` by default.
## 81                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 82                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       Arguments `Q` and `stop` are not yet provided as they depend on the task.
## 83                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 84                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 All settings are passed directly, rather than through `xgboost`'s `params` argument. `nrounds` has been set to `1` and `verbose` to `0` by default. `num_class` is set internally, so do not set this manually.
## 85                                                                                                                                                                                                                                                                                                                                                                                                                                    The `predict` method uses `cl_predict` from the `clue` package to compute the cluster memberships for new data. The default `centers = 2` is added so the method runs without setting parameters, but this must in reality of course be changed by the user.
## 86                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 87                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        A cluster index of NA indicates noise points. Specify `method = 'dist'` if the data should be interpreted as dissimilarity matrix or object. Otherwise Euclidean distances will be used.
## 88                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 89                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 90                                                                                                                                                                                                                                                                                                                                                                                                                                                                   `centers` has been set to `2L` by default. The nearest center in kernel distance determines cluster assignment of new data points. Kernel parameters have to be passed directly and not by using the `kpar` list in `kkmeans`
## 91                                                                                                                                                                                                                                                                                                                                                                                                                                    The `predict` method uses `cl_predict` from the `clue` package to compute the cluster memberships for new data. The default `centers = 2` is added so the method runs without setting parameters, but this must in reality of course be changed by the user.
## 92                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  Calls MiniBatchKmeans of package ClusterR. Argument `clusters` has default value of 2 if not provided by user.
## 93                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 94                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            You may have to install the XMeans Weka package: `WPM('install-package', 'XMeans')`.
## 95                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 96                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               `na.action` has been set to `na.impute` by default to allow missing data support.
## 97                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 98                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             `use_missing_data` has been set to `TRUE` by default to allow missing data support.
## 99                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
## 100                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 101                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 102                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 103                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 104                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Renamed parameter `learner` to `Learner` due to nameclash with `setHyperPars`. Default changes: `Learner = "ls"`, `xval = 0`, and `maxdepth = 1`.
## 105                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 106                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 107                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 108                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              See `?ctree_control` for possible breakage for nominal features with missingness.
## 109                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 110                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              See `?ctree_control` for possible breakage for nominal features with missingness.
## 111                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 112                                                                                                                                                                                                                                                                          Factors automatically get converted to dummy columns, ordered factors to integer.\n    glmnet uses a global control object for its parameters. mlr resets all control parameters to their defaults\n    before setting the specified parameters and after training.\n    If you are setting glmnet.control parameters through glmnet.control,\n    you need to save and re-set them after running the glmnet learner.
## 113                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 114                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              `pmutatemajor`, `pmutateminor`, `pcrossover`, `psplit`, and `pprune`,\n      are scaled internally to sum to 100.
## 115                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 116                                                                                                                                                                                                                                                                                                                                                                                                                                                                    Only allow one base learner for functional covariate and one base learner for scalar covariate, the parameters for these base learners are the same. Also we currently do not support interaction between scalar covariates
## 117                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 118                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 119                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 120                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 121                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 122                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 Kernel parameters have to be passed directly and not by using the `kpar` list in `gausspr`.\n    Note that `fit` has been set to `FALSE` by default for speed.
## 123                                                                                                                                                                                                                                                                                                                                                                                                                                                                                `keep.data` is set to FALSE to reduce memory requirements, `distribution` has been set to `"gaussian"` by default.Param 'n.cores' has been to set to '1' by default to suppress parallelization by the package.
## 124                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            'family' must be a character and every family has its own link, i.e. family = 'gaussian', link.gaussian = 'identity', which is also the default. We set 'model' to FALSE by default to save memory.
## 125                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 126                                                                                                         Factors automatically get converted to dummy columns, ordered factors to integer.\n      Parameter `s` (value of the regularization parameter used for predictions) is set to `0.01` by default,\n      but needs to be tuned by the user.\n      glmnet uses a global control object for its parameters. mlr resets all control parameters to their defaults\n      before setting the specified parameters and after training.\n      If you are setting glmnet.control parameters through glmnet.control,\n      you need to save and re-set them after running the glmnet learner.
## 127                                                                                                   (1) As the optimization routine assumes that the inputs are scaled to the unit hypercube [0,1]^d,\n            the input gets scaled for each variable by default. If this is not wanted, scale = FALSE has\n            to be set. (2) We replace the GPfit parameter 'corr = list(type = 'exponential',power = 1.95)' to be seperate\n            parameters 'type' and 'power', in the case of  corr = list(type = 'matern', nu = 0.5), the seperate parameters\n            are 'type' and 'matern_nu_k = 0', and nu is computed by 'nu = (2 * matern_nu_k + 1) / 2 = 0.5'\n            
## 128                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        The default value of `missing_values_handling` is `"MeanImputation"`, so missing values are automatically mean-imputed.
## 129                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 130                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                `family` is always set to `"gaussian"`. The default value of `missing_values_handling` is `"MeanImputation"`, so missing values are automatically mean-imputed.
## 131                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 132                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 133                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 134                                                                                                                                                                     In predict, we currently always use `type = "SK"`. The extra parameter `jitter` (default is `FALSE`) enables adding a very small jitter (order 1e-12) to the x-values before prediction, as `predict.km` reproduces the exact y-values of the training data points, when you pass them in, even if the nugget effect is turned on. \n We further introduced `nugget.stability` which sets the `nugget` to `nugget.stability * var(y)` before each training to improve numerical stability. We recommend a setting of 10^-8
## 135                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         Kernel parameters have to be passed directly and not by using the `kpar` list in `ksvm`. Note that `fit` has been set to `FALSE` by default for speed.
## 136                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 137                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          Parameter `svr_eps` has been set to `0.1` by default.
## 138                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     `type = 11` (the default) is primal and `type = 12` is dual problem. Parameter `svr_eps` has been set to `0.1` by default.
## 139                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 140                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 141                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 142                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         `size` has been set to `3` by default.
## 143                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 144                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 145                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      trace=FALSE was set by default to disable logging output.
## 146                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 147                                                                                                                                                                                                             See the section about 'regr.randomForest' in `?makeLearner` for information about se estimation. Note that the rf can freeze the R process if trained on a task with 1 feature which is constant. This can happen in feature forward selection, also due to resampling, and you need to remove such features with removeConstantFeatures. keep.inbag is NULL by default but if predict.type = 'se' and se.method = 'jackknife' (the default) then it is automatically set to TRUE.
## 148                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            `na.action` has been set to `"na.impute"` by default to allow missing data support.
## 149                                                                                                                                                                                                  By default, internal parallelization is switched off (`num.threads = 1`), `verbose` output is disabled, `respect.unordered.factors` is set to `order` for all splitrules. All settings are changeable. `mtry.perc` sets `mtry` to `mtry.perc*getTaskNFeats(.task)`. Default for `mtry` is the floor of square root of number of features in task. SE estimation is mc bias-corrected jackknife after bootstrap, see the section about 'regr.randomForest' in `?makeLearner` for more details.
## 150                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 151                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               `xval` has been set to `0` by default for speed.
## 152                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 153                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  You select the order of the regression by using `modelfun = "FO"` (first order), `"TWI"` (two-way interactions, this is with 1st oder terms!) and `"SO"` (full second order).
## 154                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          Kernel parameters have to be passed directly and not by using the `kpar` list in `rvm`. Note that `fit` has been set to `FALSE` by default for speed.
## 155                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 156                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            All settings are passed directly, rather than through `xgboost`'s `params` argument. `nrounds` has been set to `1` and `verbose` to `0` by default.
## 157                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              See `?ctree_control` for possible breakage for nominal features with missingness.
## 158                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
## 159                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Factors automatically get converted to dummy columns, ordered factors to integer.
## 160                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 `family` has been set to `CoxPH()` by default.
## 161                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     `keep.data` is set to FALSE to reduce memory requirements.
## 162                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 `family` has been set to `CoxPH()` by default.
## 163                                                                                                                                                       Factors automatically get converted to dummy columns, ordered factors to integer.Parameter `s` (value of the regularization parameter used for predictions) is set to `0.1` by default, but needs to be tuned by the user. glmnet uses a global control object for its parameters. mlr resets all control parameters to their defaults before setting the specified parametersand after training. If you are setting glmnet.control parameters through glmnet.control,you need to save and re-set them after running the glmnet learner.
## 164                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            `na.action` has been set to `"na.impute"` by default to allow missing data support.
## 165                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         By default, internal parallelization is switched off (`num.threads = 1`), `verbose` output is disabled, `respect.unordered.factors` is set to `order` for all splitrules. All settings are changeable.
## 166                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               `xval` has been set to `0` by default for speed.
##           type installed numerics factors ordered missings weights  prob
## 1      classif     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 2      classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 3      classif     FALSE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 4      classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 5      classif     FALSE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 6      classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 7      classif     FALSE     TRUE    TRUE   FALSE     TRUE    TRUE  TRUE
## 8      classif     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE  TRUE
## 9      classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 10     classif     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE  TRUE
## 11     classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 12     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 13     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 14     classif     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 15     classif     FALSE     TRUE    TRUE    TRUE    FALSE    TRUE  TRUE
## 16     classif     FALSE     TRUE   FALSE   FALSE    FALSE    TRUE  TRUE
## 17     classif     FALSE    FALSE   FALSE   FALSE    FALSE   FALSE  TRUE
## 18     classif     FALSE    FALSE   FALSE   FALSE    FALSE   FALSE  TRUE
## 19     classif     FALSE    FALSE   FALSE   FALSE    FALSE    TRUE  TRUE
## 20     classif     FALSE    FALSE   FALSE   FALSE    FALSE   FALSE  TRUE
## 21     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 22     classif      TRUE     TRUE    TRUE    TRUE     TRUE   FALSE  TRUE
## 23     classif     FALSE    FALSE   FALSE   FALSE    FALSE   FALSE  TRUE
## 24     classif      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 25     classif     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 26     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 27     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 28     classif      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE  TRUE
## 29     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 30     classif     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 31     classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 32     classif      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE  TRUE
## 33     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 34     classif      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE  TRUE
## 35     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 36     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 37     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 38     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 39     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 40     classif      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 41     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 42     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 43     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 44     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 45     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 46     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 47     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 48     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 49     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 50     classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 51     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 52     classif      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 53     classif     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 54     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 55     classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 56     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 57     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 58     classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 59     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 60     classif     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 61     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 62     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 63     classif      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE  TRUE
## 64     classif     FALSE     TRUE    TRUE    TRUE    FALSE   FALSE  TRUE
## 65     classif     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 66     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 67     classif      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE  TRUE
## 68     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 69     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 70     classif      TRUE     TRUE    TRUE    TRUE    FALSE   FALSE  TRUE
## 71     classif     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE  TRUE
## 72     classif      TRUE     TRUE    TRUE    TRUE    FALSE    TRUE  TRUE
## 73     classif     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 74     classif     FALSE     TRUE    TRUE    TRUE    FALSE   FALSE FALSE
## 75     classif     FALSE     TRUE   FALSE    TRUE    FALSE   FALSE FALSE
## 76     classif     FALSE     TRUE    TRUE    TRUE    FALSE   FALSE  TRUE
## 77     classif      TRUE     TRUE    TRUE    TRUE     TRUE    TRUE  TRUE
## 78     classif     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 79     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 80     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 81     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 82     classif     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 83     classif      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE  TRUE
## 84     classif      TRUE     TRUE   FALSE   FALSE     TRUE    TRUE  TRUE
## 85     cluster     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 86     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 87     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 88     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 89     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 90     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 91     cluster     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 92     cluster     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE  TRUE
## 93     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 94     cluster      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 95  multilabel     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE  TRUE
## 96  multilabel     FALSE     TRUE    TRUE   FALSE     TRUE    TRUE  TRUE
## 97  multilabel     FALSE     TRUE    TRUE    TRUE    FALSE   FALSE FALSE
## 98        regr     FALSE     TRUE    TRUE   FALSE     TRUE   FALSE FALSE
## 99        regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 100       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 101       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 102       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 103       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 104       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 105       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 106       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 107       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 108       regr     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
## 109       regr     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 110       regr     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
## 111       regr     FALSE     TRUE    TRUE   FALSE     TRUE   FALSE FALSE
## 112       regr      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 113       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 114       regr     FALSE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 115       regr     FALSE     TRUE   FALSE   FALSE    FALSE    TRUE FALSE
## 116       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 117       regr      TRUE     TRUE    TRUE    TRUE     TRUE   FALSE FALSE
## 118       regr     FALSE    FALSE   FALSE   FALSE    FALSE   FALSE FALSE
## 119       regr      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 120       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 121       regr     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 122       regr      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 123       regr      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE FALSE
## 124       regr      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 125       regr     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 126       regr      TRUE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 127       regr      TRUE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 128       regr      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE FALSE
## 129       regr      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE FALSE
## 130       regr      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE FALSE
## 131       regr      TRUE     TRUE    TRUE   FALSE     TRUE   FALSE FALSE
## 132       regr      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 133       regr      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 134       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 135       regr      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 136       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 137       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 138       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 139       regr      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 140       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 141       regr     FALSE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 142       regr      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 143       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 144       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 145       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 146       regr     FALSE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 147       regr      TRUE     TRUE    TRUE    TRUE    FALSE   FALSE FALSE
## 148       regr     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
## 149       regr      TRUE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 150       regr     FALSE     TRUE   FALSE    TRUE    FALSE   FALSE FALSE
## 151       regr      TRUE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
## 152       regr     FALSE     TRUE    TRUE    TRUE    FALSE   FALSE FALSE
## 153       regr     FALSE     TRUE   FALSE   FALSE    FALSE   FALSE FALSE
## 154       regr      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 155       regr      TRUE     TRUE    TRUE   FALSE    FALSE   FALSE FALSE
## 156       regr      TRUE     TRUE   FALSE   FALSE     TRUE    TRUE FALSE
## 157       surv     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
## 158       surv      TRUE     TRUE    TRUE   FALSE    FALSE    TRUE FALSE
## 159       surv      TRUE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 160       surv     FALSE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 161       surv      TRUE     TRUE    TRUE   FALSE     TRUE    TRUE FALSE
## 162       surv     FALSE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 163       surv      TRUE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 164       surv     FALSE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
## 165       surv      TRUE     TRUE    TRUE    TRUE    FALSE    TRUE FALSE
## 166       surv      TRUE     TRUE    TRUE    TRUE     TRUE    TRUE FALSE
##     oneclass twoclass multiclass class.weights featimp oobpreds functionals
## 1      FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 2      FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 3      FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 4      FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 5      FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 6      FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 7      FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 8      FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 9      FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 10     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 11     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 12     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 13     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 14     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 15     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 16     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 17     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE        TRUE
## 18     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 19     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 20     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 21     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE        TRUE
## 22     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE        TRUE
## 23     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE        TRUE
## 24     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 25     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 26     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 27     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 28     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 29     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 30     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 31     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 32     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 33     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 34     FALSE     TRUE      FALSE         FALSE    TRUE    FALSE       FALSE
## 35     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 36     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 37     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 38     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 39     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 40     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 41     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 42     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 43     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 44     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 45     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 46     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 47     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 48     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 49     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 50     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 51     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 52     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 53     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 54     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 55     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 56     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 57     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 58     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 59     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 60     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 61     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 62     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 63     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 64     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 65     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 66     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 67     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 68     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 69     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 70     FALSE     TRUE       TRUE          TRUE    TRUE     TRUE       FALSE
## 71     FALSE     TRUE       TRUE         FALSE    TRUE     TRUE       FALSE
## 72     FALSE     TRUE       TRUE         FALSE    TRUE     TRUE       FALSE
## 73     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 74     FALSE     TRUE       TRUE         FALSE   FALSE     TRUE       FALSE
## 75     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 76     FALSE     TRUE      FALSE         FALSE   FALSE    FALSE       FALSE
## 77     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 78     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 79     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 80     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 81     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 82     FALSE     TRUE       TRUE         FALSE   FALSE    FALSE       FALSE
## 83     FALSE     TRUE       TRUE          TRUE   FALSE    FALSE       FALSE
## 84     FALSE     TRUE       TRUE         FALSE    TRUE    FALSE       FALSE
## 85     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 86     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 87     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 88     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 89     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 90     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 91     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 92     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 93     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 94     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 95     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 96     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 97     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 98     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 99     FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 100    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 101    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 102    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 103    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 104    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 105    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 106    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 107    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 108    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 109    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 110    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 111    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 112    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 113    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 114    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 115    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 116    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE        TRUE
## 117    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE        TRUE
## 118    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE        TRUE
## 119    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 120    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 121    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 122    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 123    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 124    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 125    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 126    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 127    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 128    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 129    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 130    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 131    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 132    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 133    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 134    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 135    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 136    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 137    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 138    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 139    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 140    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 141    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 142    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 143    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 144    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 145    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 146    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 147    FALSE    FALSE      FALSE         FALSE    TRUE     TRUE       FALSE
## 148    FALSE    FALSE      FALSE         FALSE    TRUE     TRUE       FALSE
## 149    FALSE    FALSE      FALSE         FALSE    TRUE     TRUE       FALSE
## 150    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 151    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 152    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 153    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 154    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 155    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 156    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 157    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 158    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 159    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 160    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 161    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 162    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 163    FALSE    FALSE      FALSE         FALSE   FALSE    FALSE       FALSE
## 164    FALSE    FALSE      FALSE         FALSE    TRUE     TRUE       FALSE
## 165    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
## 166    FALSE    FALSE      FALSE         FALSE    TRUE    FALSE       FALSE
##     single.functional    se lcens rcens icens
## 1               FALSE FALSE FALSE FALSE FALSE
## 2               FALSE FALSE FALSE FALSE FALSE
## 3               FALSE FALSE FALSE FALSE FALSE
## 4               FALSE FALSE FALSE FALSE FALSE
## 5               FALSE FALSE FALSE FALSE FALSE
## 6               FALSE FALSE FALSE FALSE FALSE
## 7               FALSE FALSE FALSE FALSE FALSE
## 8               FALSE FALSE FALSE FALSE FALSE
## 9               FALSE FALSE FALSE FALSE FALSE
## 10              FALSE FALSE FALSE FALSE FALSE
## 11              FALSE FALSE FALSE FALSE FALSE
## 12              FALSE FALSE FALSE FALSE FALSE
## 13              FALSE FALSE FALSE FALSE FALSE
## 14              FALSE FALSE FALSE FALSE FALSE
## 15              FALSE FALSE FALSE FALSE FALSE
## 16              FALSE FALSE FALSE FALSE FALSE
## 17              FALSE FALSE FALSE FALSE FALSE
## 18               TRUE FALSE FALSE FALSE FALSE
## 19               TRUE FALSE FALSE FALSE FALSE
## 20               TRUE FALSE FALSE FALSE FALSE
## 21              FALSE FALSE FALSE FALSE FALSE
## 22              FALSE FALSE FALSE FALSE FALSE
## 23               TRUE FALSE FALSE FALSE FALSE
## 24              FALSE FALSE FALSE FALSE FALSE
## 25              FALSE FALSE FALSE FALSE FALSE
## 26              FALSE FALSE FALSE FALSE FALSE
## 27              FALSE FALSE FALSE FALSE FALSE
## 28              FALSE FALSE FALSE FALSE FALSE
## 29              FALSE FALSE FALSE FALSE FALSE
## 30              FALSE FALSE FALSE FALSE FALSE
## 31              FALSE FALSE FALSE FALSE FALSE
## 32              FALSE FALSE FALSE FALSE FALSE
## 33              FALSE FALSE FALSE FALSE FALSE
## 34              FALSE FALSE FALSE FALSE FALSE
## 35              FALSE FALSE FALSE FALSE FALSE
## 36              FALSE FALSE FALSE FALSE FALSE
## 37              FALSE FALSE FALSE FALSE FALSE
## 38              FALSE FALSE FALSE FALSE FALSE
## 39              FALSE FALSE FALSE FALSE FALSE
## 40              FALSE FALSE FALSE FALSE FALSE
## 41              FALSE FALSE FALSE FALSE FALSE
## 42              FALSE FALSE FALSE FALSE FALSE
## 43              FALSE FALSE FALSE FALSE FALSE
## 44              FALSE FALSE FALSE FALSE FALSE
## 45              FALSE FALSE FALSE FALSE FALSE
## 46              FALSE FALSE FALSE FALSE FALSE
## 47              FALSE FALSE FALSE FALSE FALSE
## 48              FALSE FALSE FALSE FALSE FALSE
## 49              FALSE FALSE FALSE FALSE FALSE
## 50              FALSE FALSE FALSE FALSE FALSE
## 51              FALSE FALSE FALSE FALSE FALSE
## 52              FALSE FALSE FALSE FALSE FALSE
## 53              FALSE FALSE FALSE FALSE FALSE
## 54              FALSE FALSE FALSE FALSE FALSE
## 55              FALSE FALSE FALSE FALSE FALSE
## 56              FALSE FALSE FALSE FALSE FALSE
## 57              FALSE FALSE FALSE FALSE FALSE
## 58              FALSE FALSE FALSE FALSE FALSE
## 59              FALSE FALSE FALSE FALSE FALSE
## 60              FALSE FALSE FALSE FALSE FALSE
## 61              FALSE FALSE FALSE FALSE FALSE
## 62              FALSE FALSE FALSE FALSE FALSE
## 63              FALSE FALSE FALSE FALSE FALSE
## 64              FALSE FALSE FALSE FALSE FALSE
## 65              FALSE FALSE FALSE FALSE FALSE
## 66              FALSE FALSE FALSE FALSE FALSE
## 67              FALSE FALSE FALSE FALSE FALSE
## 68              FALSE FALSE FALSE FALSE FALSE
## 69              FALSE FALSE FALSE FALSE FALSE
## 70              FALSE FALSE FALSE FALSE FALSE
## 71              FALSE FALSE FALSE FALSE FALSE
## 72              FALSE FALSE FALSE FALSE FALSE
## 73              FALSE FALSE FALSE FALSE FALSE
## 74              FALSE FALSE FALSE FALSE FALSE
## 75              FALSE FALSE FALSE FALSE FALSE
## 76              FALSE FALSE FALSE FALSE FALSE
## 77              FALSE FALSE FALSE FALSE FALSE
## 78              FALSE FALSE FALSE FALSE FALSE
## 79              FALSE FALSE FALSE FALSE FALSE
## 80              FALSE FALSE FALSE FALSE FALSE
## 81              FALSE FALSE FALSE FALSE FALSE
## 82              FALSE FALSE FALSE FALSE FALSE
## 83              FALSE FALSE FALSE FALSE FALSE
## 84              FALSE FALSE FALSE FALSE FALSE
## 85              FALSE FALSE FALSE FALSE FALSE
## 86              FALSE FALSE FALSE FALSE FALSE
## 87              FALSE FALSE FALSE FALSE FALSE
## 88              FALSE FALSE FALSE FALSE FALSE
## 89              FALSE FALSE FALSE FALSE FALSE
## 90              FALSE FALSE FALSE FALSE FALSE
## 91              FALSE FALSE FALSE FALSE FALSE
## 92              FALSE FALSE FALSE FALSE FALSE
## 93              FALSE FALSE FALSE FALSE FALSE
## 94              FALSE FALSE FALSE FALSE FALSE
## 95              FALSE FALSE FALSE FALSE FALSE
## 96              FALSE FALSE FALSE FALSE FALSE
## 97              FALSE FALSE FALSE FALSE FALSE
## 98              FALSE FALSE FALSE FALSE FALSE
## 99              FALSE  TRUE FALSE FALSE FALSE
## 100             FALSE  TRUE FALSE FALSE FALSE
## 101             FALSE  TRUE FALSE FALSE FALSE
## 102             FALSE  TRUE FALSE FALSE FALSE
## 103             FALSE FALSE FALSE FALSE FALSE
## 104             FALSE FALSE FALSE FALSE FALSE
## 105             FALSE  TRUE FALSE FALSE FALSE
## 106             FALSE  TRUE FALSE FALSE FALSE
## 107             FALSE  TRUE FALSE FALSE FALSE
## 108             FALSE FALSE FALSE FALSE FALSE
## 109             FALSE  TRUE FALSE FALSE FALSE
## 110             FALSE FALSE FALSE FALSE FALSE
## 111             FALSE FALSE FALSE FALSE FALSE
## 112             FALSE FALSE FALSE FALSE FALSE
## 113             FALSE FALSE FALSE FALSE FALSE
## 114             FALSE FALSE FALSE FALSE FALSE
## 115             FALSE FALSE FALSE FALSE FALSE
## 116             FALSE FALSE FALSE FALSE FALSE
## 117             FALSE FALSE FALSE FALSE FALSE
## 118              TRUE FALSE FALSE FALSE FALSE
## 119             FALSE FALSE FALSE FALSE FALSE
## 120             FALSE FALSE FALSE FALSE FALSE
## 121             FALSE FALSE FALSE FALSE FALSE
## 122             FALSE  TRUE FALSE FALSE FALSE
## 123             FALSE FALSE FALSE FALSE FALSE
## 124             FALSE  TRUE FALSE FALSE FALSE
## 125             FALSE FALSE FALSE FALSE FALSE
## 126             FALSE FALSE FALSE FALSE FALSE
## 127             FALSE  TRUE FALSE FALSE FALSE
## 128             FALSE FALSE FALSE FALSE FALSE
## 129             FALSE FALSE FALSE FALSE FALSE
## 130             FALSE FALSE FALSE FALSE FALSE
## 131             FALSE FALSE FALSE FALSE FALSE
## 132             FALSE FALSE FALSE FALSE FALSE
## 133             FALSE FALSE FALSE FALSE FALSE
## 134             FALSE  TRUE FALSE FALSE FALSE
## 135             FALSE FALSE FALSE FALSE FALSE
## 136             FALSE  TRUE FALSE FALSE FALSE
## 137             FALSE FALSE FALSE FALSE FALSE
## 138             FALSE FALSE FALSE FALSE FALSE
## 139             FALSE  TRUE FALSE FALSE FALSE
## 140             FALSE FALSE FALSE FALSE FALSE
## 141             FALSE FALSE FALSE FALSE FALSE
## 142             FALSE FALSE FALSE FALSE FALSE
## 143             FALSE FALSE FALSE FALSE FALSE
## 144             FALSE FALSE FALSE FALSE FALSE
## 145             FALSE FALSE FALSE FALSE FALSE
## 146             FALSE FALSE FALSE FALSE FALSE
## 147             FALSE  TRUE FALSE FALSE FALSE
## 148             FALSE FALSE FALSE FALSE FALSE
## 149             FALSE  TRUE FALSE FALSE FALSE
## 150             FALSE FALSE FALSE FALSE FALSE
## 151             FALSE FALSE FALSE FALSE FALSE
## 152             FALSE FALSE FALSE FALSE FALSE
## 153             FALSE FALSE FALSE FALSE FALSE
## 154             FALSE FALSE FALSE FALSE FALSE
## 155             FALSE FALSE FALSE FALSE FALSE
## 156             FALSE FALSE FALSE FALSE FALSE
## 157             FALSE FALSE FALSE FALSE FALSE
## 158             FALSE FALSE FALSE FALSE FALSE
## 159             FALSE FALSE FALSE FALSE FALSE
## 160             FALSE FALSE FALSE FALSE FALSE
## 161             FALSE FALSE FALSE FALSE FALSE
## 162             FALSE FALSE FALSE FALSE FALSE
## 163             FALSE FALSE FALSE FALSE FALSE
## 164             FALSE FALSE FALSE FALSE FALSE
## 165             FALSE FALSE FALSE FALSE FALSE
## 166             FALSE FALSE FALSE FALSE FALSE
(dt_task <- makeClassifTask(data = train, target = "diabetes"))
## Supervised task: train
## Type: classif
## Target: diabetes
## Observations: 614
## Features:
##    numerics     factors     ordered functionals 
##           8           0           0           0 
## Missings: FALSE
## Has weights: FALSE
## Has blocking: FALSE
## Has coordinates: FALSE
## Classes: 2
## neg pos 
## 386 228 
## Positive class: neg
(dt_prob <- makeLearner('classif.gbm', predict.type = "prob"))
## Learner classif.gbm from package gbm
## Type: classif
## Name: Gradient Boosting Machine; Short name: gbm
## Class: classif.gbm
## Properties: twoclass,multiclass,missings,numerics,factors,prob,weights,featimp
## Predict-Type: prob
## Hyperparameters: keep.data=FALSE

Feature Selection

library(FSelector)
listFilterMethods()
##                                       id         package
## 1                             anova.test                
## 2                                    auc                
## 3                               carscore            care
## 4                  FSelector_chi.squared       FSelector
## 5                   FSelector_gain.ratio       FSelector
## 6             FSelector_information.gain       FSelector
## 7                         FSelector_oneR       FSelector
## 8                       FSelector_relief       FSelector
## 9      FSelector_symmetrical.uncertainty       FSelector
## 10              FSelectorRcpp_gain.ratio   FSelectorRcpp
## 11        FSelectorRcpp_information.gain   FSelectorRcpp
## 12                  FSelectorRcpp_relief   FSelectorRcpp
## 13 FSelectorRcpp_symmetrical.uncertainty   FSelectorRcpp
## 14                          kruskal.test                
## 15                    linear.correlation                
## 16                                  mrmr           mRMRe
## 17              party_cforest.importance           party
## 18                permutation.importance                
## 19                          praznik_CMIM         praznik
## 20                          praznik_DISR         praznik
## 21                           praznik_JMI         praznik
## 22                          praznik_JMIM         praznik
## 23                           praznik_MIM         praznik
## 24                          praznik_MRMR         praznik
## 25                         praznik_NJMIM         praznik
## 26               randomForest_importance    randomForest
## 27            randomForestSRC_importance randomForestSRC
## 28            randomForestSRC_var.select randomForestSRC
## 29                       ranger_impurity          ranger
## 30                    ranger_permutation          ranger
## 31                      rank.correlation                
## 32                univariate.model.score                
## 33                              variance                
##                                        desc
## 1  ANOVA Test for binary and multiclass ...
## 2  AUC filter for binary classification ...
## 3                                CAR scores
## 4  Chi-squared statistic of independence...
## 5  Chi-squared statistic of independence...
## 6  Entropy-based information gain betwee...
## 7                     oneR association rule
## 8                          RELIEF algorithm
## 9  Entropy-based symmetrical uncertainty...
## 10 Entropy-based Filters: Algorithms tha...
## 11 Entropy-based Filters: Algorithms tha...
## 12                         RELIEF algorithm
## 13 Entropy-based Filters: Algorithms tha...
## 14 Kruskal Test for binary and multiclas...
## 15 Pearson correlation between feature a...
## 16 Minimum redundancy, maximum relevance...
## 17 Permutation importance of random fore...
## 18 Aggregated difference between feature...
## 19 Minimal conditional mutual informatio...
## 20 Double input symmetrical relevance fi...
## 21          Joint mutual information filter
## 22 Minimal joint mutual information maxi...
## 23 conditional mutual information based ...
## 24 Minimum redundancy maximal relevancy ...
## 25 Minimal normalised joint mutual infor...
## 26 Importance based on OOB-accuracy or n...
## 27 Importance of random forests fitted i...
## 28 Minimal depth of / variable hunting v...
## 29 Variable importance based on ranger i...
## 30 Variable importance based on ranger p...
## 31 Spearman's correlation between featur...
## 32 Resamples an mlr learner for each inp...
## 33                 A simple variance filter
listFilterEnsembleMethods()
##         id
## 1  E-Borda
## 2    E-max
## 3   E-mean
## 4 E-median
## 5    E-min
##                                                                                                     desc
## 1                  Borda ensemble filter. Takes the sum across all base filter methods for each feature.
## 2 Maximum ensemble filter. Takes the best maximum value across all base filter methods for each feature.
## 3                  Mean ensemble filter. Takes the mean across all base filter methods for each feature.
## 4              Median ensemble filter. Takes the median across all base filter methods for each feature.
## 5 Minimum ensemble filter. Takes the best minimum value across all base filter methods for each feature.
generateFilterValuesData(dt_task, method = "FSelector_information.gain") %>% 
  plotFilterValues() +
  theme_bw() +
    labs(x = "feature",
         y = "information gain",
         title = "Information gain of features in GBM",
         caption = "Source: Pima Indians Diabetes Database")

feat_imp_tpr <- generateFeatureImportanceData(task = dt_task, 
                              learner = dt_prob,
                              measure = tpr, 
                              interaction = FALSE)
## Distribution not specified, assuming bernoulli ...
feat_imp_tpr$res %>%
  gather() %>%
  ggplot(aes(x = reorder(key, value), y = value)) +
    geom_bar(stat = "identity") +
    labs(x = "feature",
         title = "True positive rate of features in GBM",
         subtitle = "calculated with permutation importance",
         caption = "Source: Pima Indians Diabetes Database")

feat_imp_auc <- generateFeatureImportanceData(task = dt_task, 
                              learner = dt_prob,
                              measure = auc, 
                              interaction = FALSE)
## Distribution not specified, assuming bernoulli ...
feat_imp_auc$res %>%
  gather() %>%
  ggplot(aes(x = reorder(key, value), y = value)) +
    geom_bar(stat = "identity") +
    labs(x = "feature",
         title = "Area under the curve of features in GBM",
         subtitle = "calculated with permutation importance",
         caption = "Source: Pima Indians Diabetes Database")

set.seed(1000) 
train <- dplyr::select(train, -pedigree, -pressure, -triceps) 
test <- dplyr::select(test, -pedigree, -pressure, -triceps)
list( train = summary(train), test = summary(test) )
## $train
##     pregnant         glucose         insulin            mass      
##  Min.   : 0.000   Min.   :  0.0   Min.   :  0.00   Min.   : 0.00  
##  1st Qu.: 1.000   1st Qu.:100.0   1st Qu.:  0.00   1st Qu.:27.10  
##  Median : 3.000   Median :119.0   Median : 36.50   Median :32.00  
##  Mean   : 3.894   Mean   :123.1   Mean   : 81.65   Mean   :31.92  
##  3rd Qu.: 6.000   3rd Qu.:143.0   3rd Qu.:131.50   3rd Qu.:36.38  
##  Max.   :17.000   Max.   :199.0   Max.   :846.00   Max.   :59.40  
##       age        diabetes 
##  Min.   :21.00   neg:386  
##  1st Qu.:24.00   pos:228  
##  Median :29.00            
##  Mean   :33.42            
##  3rd Qu.:41.00            
##  Max.   :81.00            
## 
## $test
##     pregnant         glucose         insulin           mass      
##  Min.   : 0.000   Min.   :  0.0   Min.   :  0.0   Min.   : 0.00  
##  1st Qu.: 1.000   1st Qu.: 93.0   1st Qu.:  0.0   1st Qu.:27.80  
##  Median : 2.000   Median :108.0   Median : 20.5   Median :32.40  
##  Mean   : 3.649   Mean   :112.3   Mean   : 72.4   Mean   :32.29  
##  3rd Qu.: 6.000   3rd Qu.:133.8   3rd Qu.:100.0   3rd Qu.:36.88  
##  Max.   :14.000   Max.   :197.0   Max.   :744.0   Max.   :67.10  
##       age        diabetes 
##  Min.   :21.00   neg:114  
##  1st Qu.:23.25   pos: 40  
##  Median :29.00            
##  Mean   :32.54            
##  3rd Qu.:39.75            
##  Max.   :67.00
(dt_task <- makeClassifTask(data = train, target = "diabetes"))
## Supervised task: train
## Type: classif
## Target: diabetes
## Observations: 614
## Features:
##    numerics     factors     ordered functionals 
##           5           0           0           0 
## Missings: FALSE
## Has weights: FALSE
## Has blocking: FALSE
## Has coordinates: FALSE
## Classes: 2
## neg pos 
## 386 228 
## Positive class: neg

Hyperparameter Optimization

getParamSet("classif.gbm")
##                       Type len       Def
## distribution      discrete   - bernoulli
## n.trees            integer   -       100
## cv.folds           integer   -         0
## interaction.depth  integer   -         1
## n.minobsinnode     integer   -        10
## shrinkage          numeric   -       0.1
## bag.fraction       numeric   -       0.5
## train.fraction     numeric   -         1
## keep.data          logical   -      TRUE
## verbose            logical   -     FALSE
## n.cores            integer   -         1
##                                                     Constr Req Tunable Trafo
## distribution      gaussian,bernoulli,huberized,adaboost...   -    TRUE     -
## n.trees                                           1 to Inf   -    TRUE     -
## cv.folds                                       -Inf to Inf   -    TRUE     -
## interaction.depth                                 1 to Inf   -    TRUE     -
## n.minobsinnode                                    1 to Inf   -    TRUE     -
## shrinkage                                         0 to Inf   -    TRUE     -
## bag.fraction                                        0 to 1   -    TRUE     -
## train.fraction                                      0 to 1   -    TRUE     -
## keep.data                                                -   -   FALSE     -
## verbose                                                  -   -   FALSE     -
## n.cores                                        -Inf to Inf   -   FALSE     -
dt_param <- makeParamSet( 
  makeIntegerParam("n.trees", lower = 20, upper = 150),
  makeNumericParam("shrinkage", lower = 0.01, upper = 0.1))

ctrl = makeTuneControlGrid()

rdesc = makeResampleDesc("CV", 
                         iters = 3L, 
                         stratify = TRUE)
set.seed(1000) 
(dt_tuneparam <- tuneParams(learner = dt_prob, 
                             resampling = rdesc, 
                             measures = list(tpr,auc, fnr, mmce, tnr, setAggregation(tpr, test.sd)), 
                             par.set = dt_param, 
                             control = ctrl, 
                             task = dt_task, 
                             show.info = TRUE))
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Distribution not specified, assuming bernoulli ...
## Tune result:
## Op. pars: n.trees=20; shrinkage=0.02
## tpr.test.mean=1.0000000,auc.test.mean=0.7878691,fnr.test.mean=0.0000000,mmce.test.mean=0.3713375,tnr.test.mean=0.0000000,tpr.test.sd=0.0000000
data = generateHyperParsEffectData(dt_tuneparam, 
                                   partial.dep = TRUE)

plotHyperParsEffect(data, x = "n.trees", y = "tpr.test.mean", partial.dep.learn = makeLearner("regr.gbm"))

plotHyperParsEffect(data, x = "shrinkage", y = "tpr.test.mean", partial.dep.learn = makeLearner("regr.gbm"))

plotHyperParsEffect(data, 
                    x = "n.trees", 
                    y = "shrinkage",
                    z = "tpr.test.mean", 
                    plot.type = "heatmap",
                    partial.dep.learn = makeLearner("regr.gbm")) +
  theme_bw() +
    labs(title = "Hyperparameter effects data",
         subtitle = "of GBM model with reduced feature set",
         caption = "Source: Pima Indians Diabetes Database")

list( `Optimal HyperParameters` = dt_tuneparam$x, 
      `Optimal Metrics` = dt_tuneparam$y )
## $`Optimal HyperParameters`
## $`Optimal HyperParameters`$n.trees
## [1] 20
## 
## $`Optimal HyperParameters`$shrinkage
## [1] 0.02
## 
## 
## $`Optimal Metrics`
##  tpr.test.mean  auc.test.mean  fnr.test.mean mmce.test.mean  tnr.test.mean 
##      1.0000000      0.7878691      0.0000000      0.3713375      0.0000000 
##    tpr.test.sd 
##      0.0000000
gbm_final <- setHyperPars(dt_prob, par.vals = dt_tuneparam$x)

set.seed(1000) 
gbm_final_train <- train(learner = gbm_final, task = dt_task) 
## Distribution not specified, assuming bernoulli ...
getLearnerModel(gbm_final_train)
## gbm::gbm(formula = f, data = d, n.trees = 20L, shrinkage = 0.02, 
##     keep.data = FALSE)
## A gradient boosted model with bernoulli loss function.
## 20 iterations were performed.
## There were 5 predictors of which 3 had non-zero influence.

Decision Trees

library(rpart)
library(rpart.plot)

rpart_tree <- rpart(diabetes ~ .,
                    data = train,
                    method = "class")
rpart.plot(rpart_tree, 
           roundint=FALSE, 
           type = 3, 
           clip.right.labs = FALSE)

rpart.rules(rpart_tree, roundint = FALSE)
##  diabetes                                                                           
##      0.05 when glucose <  128        & mass <  27       & age >= 29                 
##      0.10 when glucose <  128                           & age <  29                 
##      0.17 when glucose is 128 to 146 & mass <  30                                   
##      0.25 when glucose >=        146 & mass <  30       & age <  29                 
##      0.28 when glucose <  128        & mass >=       29 & age >= 29 & insulin <  143
##      0.38 when glucose is 128 to 158 & mass is 32 to 42 & age <  43                 
##      0.62 when glucose >=        146 & mass <  30       & age >= 29                 
##      0.63 when glucose <  128        & mass is 27 to 29 & age >= 29 & insulin <  143
##      0.77 when glucose <  128        & mass >=       27 & age >= 29 & insulin >= 143
##      0.82 when glucose is 128 to 158 & mass >=       42 & age <  43                 
##      0.86 when glucose is 128 to 158 & mass >=       30 & age >= 43                 
##      0.86 when glucose >=        158 & mass >=       30                             
##      0.88 when glucose is 128 to 158 & mass is 30 to 32 & age <  43

Prediction

set.seed(1000) 
(gbm_final_predict <- predict(gbm_final_train, newdata = test))
## Prediction: 154 observations
## predict.type: prob
## threshold: neg=0.50,pos=0.50
## time: 0.00
##    truth  prob.pos  prob.neg response
## 12   pos 0.4807717 0.5192283      neg
## 18   pos 0.3229851 0.6770149      neg
## 19   neg 0.3229851 0.6770149      neg
## 20   pos 0.3300235 0.6699765      neg
## 34   neg 0.3091184 0.6908816      neg
## 38   pos 0.3229851 0.6770149      neg
## ... (#rows: 154, #cols: 4)
gbm_final_predict %>% calculateROCMeasures()
##      predicted
## true  neg       pos                        
##   neg 114       0        tpr: 1   fnr: 0   
##   pos 40        0        fpr: 1   tnr: 0   
##       ppv: 0.74 for: NaN lrp: 1   acc: 0.74
##       fdr: 0.26 npv: NaN lrm: NaN dor: NaN 
## 
## 
## Abbreviations:
## tpr - True positive rate (Sensitivity, Recall)
## fpr - False positive rate (Fall-out)
## fnr - False negative rate (Miss rate)
## tnr - True negative rate (Specificity)
## ppv - Positive predictive value (Precision)
## for - False omission rate
## lrp - Positive likelihood ratio (LR+)
## fdr - False discovery rate
## npv - Negative predictive value
## acc - Accuracy
## lrm - Negative likelihood ratio (LR-)
## dor - Diagnostic odds ratio
model_performance <- performance(gbm_final_predict, 
                                 measures = list(tpr, auc, mmce, acc, tnr)) %>% 
  as.data.frame(row.names = c("True Positive Rate","Area Under Curve", "Mean Misclassification Error","Accuracy","True Negative Rate")) 

model_performance
##                                      .
## True Positive Rate           1.0000000
## Area Under Curve             0.7695175
## Mean Misclassification Error 0.2597403
## Accuracy                     0.7402597
## True Negative Rate           0.0000000
gbm_final_threshold <- generateThreshVsPerfData(gbm_final_predict, 
                                                 measures = list(tpr, auc, mmce, tnr))
gbm_final_threshold %>% 
   plotROCCurves() + 
   geom_point() +
    theme_bw() +
    labs(title = "ROC curve from predictions",
         subtitle = "of GBM model with reduced feature set",
         caption = "Source: Pima Indians Diabetes Database")

gbm_final_threshold %>% 
   plotThreshVsPerf() + 
   geom_point() +
    theme_bw() +
    labs(title = "Threshold vs. performance",
         subtitle = "for 2-class classification of GBM model with reduced feature set",
         caption = "Source: Pima Indians Diabetes Database")

gbm_final_threshold$data
##           tpr       auc      mmce   tnr  threshold
## 1   1.0000000 0.7695175 0.2597403 0.000 0.00000000
## 2   1.0000000 0.7695175 0.2597403 0.000 0.01010101
## 3   1.0000000 0.7695175 0.2597403 0.000 0.02020202
## 4   1.0000000 0.7695175 0.2597403 0.000 0.03030303
## 5   1.0000000 0.7695175 0.2597403 0.000 0.04040404
## 6   1.0000000 0.7695175 0.2597403 0.000 0.05050505
## 7   1.0000000 0.7695175 0.2597403 0.000 0.06060606
## 8   1.0000000 0.7695175 0.2597403 0.000 0.07070707
## 9   1.0000000 0.7695175 0.2597403 0.000 0.08080808
## 10  1.0000000 0.7695175 0.2597403 0.000 0.09090909
## 11  1.0000000 0.7695175 0.2597403 0.000 0.10101010
## 12  1.0000000 0.7695175 0.2597403 0.000 0.11111111
## 13  1.0000000 0.7695175 0.2597403 0.000 0.12121212
## 14  1.0000000 0.7695175 0.2597403 0.000 0.13131313
## 15  1.0000000 0.7695175 0.2597403 0.000 0.14141414
## 16  1.0000000 0.7695175 0.2597403 0.000 0.15151515
## 17  1.0000000 0.7695175 0.2597403 0.000 0.16161616
## 18  1.0000000 0.7695175 0.2597403 0.000 0.17171717
## 19  1.0000000 0.7695175 0.2597403 0.000 0.18181818
## 20  1.0000000 0.7695175 0.2597403 0.000 0.19191919
## 21  1.0000000 0.7695175 0.2597403 0.000 0.20202020
## 22  1.0000000 0.7695175 0.2597403 0.000 0.21212121
## 23  1.0000000 0.7695175 0.2597403 0.000 0.22222222
## 24  1.0000000 0.7695175 0.2597403 0.000 0.23232323
## 25  1.0000000 0.7695175 0.2597403 0.000 0.24242424
## 26  1.0000000 0.7695175 0.2597403 0.000 0.25252525
## 27  1.0000000 0.7695175 0.2597403 0.000 0.26262626
## 28  1.0000000 0.7695175 0.2597403 0.000 0.27272727
## 29  1.0000000 0.7695175 0.2597403 0.000 0.28282828
## 30  1.0000000 0.7695175 0.2597403 0.000 0.29292929
## 31  1.0000000 0.7695175 0.2597403 0.000 0.30303030
## 32  1.0000000 0.7695175 0.2597403 0.000 0.31313131
## 33  1.0000000 0.7695175 0.2597403 0.000 0.32323232
## 34  1.0000000 0.7695175 0.2597403 0.000 0.33333333
## 35  1.0000000 0.7695175 0.2597403 0.000 0.34343434
## 36  1.0000000 0.7695175 0.2597403 0.000 0.35353535
## 37  1.0000000 0.7695175 0.2597403 0.000 0.36363636
## 38  1.0000000 0.7695175 0.2597403 0.000 0.37373737
## 39  1.0000000 0.7695175 0.2597403 0.000 0.38383838
## 40  1.0000000 0.7695175 0.2597403 0.000 0.39393939
## 41  1.0000000 0.7695175 0.2597403 0.000 0.40404040
## 42  1.0000000 0.7695175 0.2597403 0.000 0.41414141
## 43  1.0000000 0.7695175 0.2597403 0.000 0.42424242
## 44  1.0000000 0.7695175 0.2597403 0.000 0.43434343
## 45  1.0000000 0.7695175 0.2597403 0.000 0.44444444
## 46  1.0000000 0.7695175 0.2597403 0.000 0.45454545
## 47  1.0000000 0.7695175 0.2597403 0.000 0.46464646
## 48  1.0000000 0.7695175 0.2597403 0.000 0.47474747
## 49  1.0000000 0.7695175 0.2597403 0.000 0.48484848
## 50  1.0000000 0.7695175 0.2597403 0.000 0.49494949
## 51  1.0000000 0.7695175 0.2597403 0.000 0.50505051
## 52  1.0000000 0.7695175 0.2597403 0.000 0.51515152
## 53  0.9912281 0.7695175 0.2142857 0.200 0.52525253
## 54  0.9824561 0.7695175 0.2012987 0.275 0.53535354
## 55  0.9736842 0.7695175 0.2077922 0.275 0.54545455
## 56  0.9298246 0.7695175 0.2207792 0.350 0.55555556
## 57  0.9210526 0.7695175 0.2207792 0.375 0.56565657
## 58  0.8771930 0.7695175 0.2467532 0.400 0.57575758
## 59  0.8157895 0.7695175 0.2792208 0.450 0.58585859
## 60  0.8070175 0.7695175 0.2727273 0.500 0.59595960
## 61  0.7807018 0.7695175 0.2857143 0.525 0.60606061
## 62  0.7807018 0.7695175 0.2857143 0.525 0.61616162
## 63  0.7807018 0.7695175 0.2857143 0.525 0.62626263
## 64  0.7807018 0.7695175 0.2857143 0.525 0.63636364
## 65  0.7456140 0.7695175 0.3051948 0.550 0.64646465
## 66  0.7456140 0.7695175 0.3051948 0.550 0.65656566
## 67  0.7280702 0.7695175 0.3116883 0.575 0.66666667
## 68  0.6491228 0.7695175 0.3311688 0.725 0.67676768
## 69  0.1666667 0.7695175 0.6168831 1.000 0.68686869
## 70  0.0000000 0.7695175 0.7402597 1.000 0.69696970
## 71  0.0000000 0.7695175 0.7402597 1.000 0.70707071
## 72  0.0000000 0.7695175 0.7402597 1.000 0.71717172
## 73  0.0000000 0.7695175 0.7402597 1.000 0.72727273
## 74  0.0000000 0.7695175 0.7402597 1.000 0.73737374
## 75  0.0000000 0.7695175 0.7402597 1.000 0.74747475
## 76  0.0000000 0.7695175 0.7402597 1.000 0.75757576
## 77  0.0000000 0.7695175 0.7402597 1.000 0.76767677
## 78  0.0000000 0.7695175 0.7402597 1.000 0.77777778
## 79  0.0000000 0.7695175 0.7402597 1.000 0.78787879
## 80  0.0000000 0.7695175 0.7402597 1.000 0.79797980
## 81  0.0000000 0.7695175 0.7402597 1.000 0.80808081
## 82  0.0000000 0.7695175 0.7402597 1.000 0.81818182
## 83  0.0000000 0.7695175 0.7402597 1.000 0.82828283
## 84  0.0000000 0.7695175 0.7402597 1.000 0.83838384
## 85  0.0000000 0.7695175 0.7402597 1.000 0.84848485
## 86  0.0000000 0.7695175 0.7402597 1.000 0.85858586
## 87  0.0000000 0.7695175 0.7402597 1.000 0.86868687
## 88  0.0000000 0.7695175 0.7402597 1.000 0.87878788
## 89  0.0000000 0.7695175 0.7402597 1.000 0.88888889
## 90  0.0000000 0.7695175 0.7402597 1.000 0.89898990
## 91  0.0000000 0.7695175 0.7402597 1.000 0.90909091
## 92  0.0000000 0.7695175 0.7402597 1.000 0.91919192
## 93  0.0000000 0.7695175 0.7402597 1.000 0.92929293
## 94  0.0000000 0.7695175 0.7402597 1.000 0.93939394
## 95  0.0000000 0.7695175 0.7402597 1.000 0.94949495
## 96  0.0000000 0.7695175 0.7402597 1.000 0.95959596
## 97  0.0000000 0.7695175 0.7402597 1.000 0.96969697
## 98  0.0000000 0.7695175 0.7402597 1.000 0.97979798
## 99  0.0000000 0.7695175 0.7402597 1.000 0.98989899
## 100 0.0000000 0.7695175 0.7402597 1.000 1.00000000
gbm_final_thr <- gbm_final_predict %>% 
  setThreshold(0.59595960) 

(dt_performance <- gbm_final_thr %>% performance(measures = list(tpr, auc, mmce, tnr)) )
##       tpr       auc      mmce       tnr 
## 0.8070175 0.7695175 0.2727273 0.5000000
(dt_cm <- gbm_final_thr %>% calculateROCMeasures() )
##      predicted
## true  neg       pos                          
##   neg 92        22        tpr: 0.81 fnr: 0.19
##   pos 20        20        fpr: 0.5  tnr: 0.5 
##       ppv: 0.82 for: 0.52 lrp: 1.61 acc: 0.73
##       fdr: 0.18 npv: 0.48 lrm: 0.39 dor: 4.18
## 
## 
## Abbreviations:
## tpr - True positive rate (Sensitivity, Recall)
## fpr - False positive rate (Fall-out)
## fnr - False negative rate (Miss rate)
## tnr - True negative rate (Specificity)
## ppv - Positive predictive value (Precision)
## for - False omission rate
## lrp - Positive likelihood ratio (LR+)
## fdr - False discovery rate
## npv - Negative predictive value
## acc - Accuracy
## lrm - Negative likelihood ratio (LR-)
## dor - Diagnostic odds ratio
performance_threshold <- performance(gbm_final_thr, measures = list(tpr, auc, mmce, acc, tnr)) %>% 
  as.data.frame(row.names = c("True Positive Rate", "Area Under Curve", "Mean Misclassification Error", "Accuracy", "True Negative Rate"))

performance_threshold
##                                      .
## True Positive Rate           0.8070175
## Area Under Curve             0.7695175
## Mean Misclassification Error 0.2727273
## Accuracy                     0.7272727
## True Negative Rate           0.5000000

Decision Boundaries

#remotes::install_github("grantmcdermott/parttree")
library(parsnip)
library(parttree)
set.seed(123) ## For consistent jitter

## Build our tree using parsnip (but with rpart as the model engine)
ti_tree =
  decision_tree() %>%
  set_engine("rpart") %>%
  set_mode("classification") %>%
  fit(diabetes ~ glucose + mass, data = PimaIndiansDiabetes)

## Plot the data and model partitions
PimaIndiansDiabetes %>%
  ggplot(aes(x = glucose, y = mass)) +
  geom_jitter(aes(col = diabetes), alpha = 0.7) +
  geom_parttree(data = ti_tree, aes(fill = diabetes), alpha = 0.1) +
  theme_bw() +
    labs(title = "Decision boundaries",
         subtitle = "for 2-class classification of RPART model (glucose + mass)",
         caption = "Source: Pima Indians Diabetes Database")

Graphical representation of a model in TensorBoard

https://www.tensorflow.org/tensorboard

Word Embeddings

The Unreasonable Effectiveness of Recurrent Neural Networks; Karpathy, 2015

Seq2Seq-Vis: Visual Debugging Tool for Sequence-to-Sequence Models; Strobelt, 2018

https://arxiv.org/pdf/1611.04558.pdf

Image classifiers are effective in practice

Visualizing and Understanding Convolutional Networks; Zeiler & Fergus, 2013

The Building Blocks of Interpretability; Olah, Satyanarayan, Johnson, Carter, Schubert, Ye, Mordvintsev

playground.tensorflow.org

Distill.pub

research.google.com/bigpicture/attacking-discrimination-in-ml

Google Creative Lab: https://quickdraw.withgoogle.com/

https://poloclub.github.io/ganlab/

http://lstm.seas.harvard.edu/


devtools::session_info()
## ─ Session info ───────────────────────────────────────────────────────────────
##  setting  value                       
##  version  R version 4.0.4 (2021-02-15)
##  os       macOS Big Sur 10.16         
##  system   x86_64, darwin17.0          
##  ui       X11                         
##  language (EN)                        
##  collate  en_US.UTF-8                 
##  ctype    en_US.UTF-8                 
##  tz       Europe/Berlin               
##  date     2021-04-11                  
## 
## ─ Packages ───────────────────────────────────────────────────────────────────
##  package      * version    date       lib
##  assertthat     0.2.1      2019-03-21 [2]
##  backports      1.2.1      2020-12-09 [2]
##  BBmisc         1.11       2017-03-10 [2]
##  broom          0.7.5      2021-02-19 [2]
##  bslib          0.2.4      2021-01-25 [2]
##  cachem         1.0.4      2021-02-13 [2]
##  callr          3.5.1      2020-10-13 [2]
##  cellranger     1.1.0      2016-07-27 [2]
##  checkmate      2.0.0      2020-02-06 [2]
##  cli            2.3.1      2021-02-23 [2]
##  colorspace     2.0-0      2020-11-11 [2]
##  crayon         1.4.1      2021-02-08 [2]
##  data.table     1.14.0     2021-02-21 [2]
##  DBI            1.1.1      2021-01-15 [2]
##  dbplyr         2.1.0      2021-02-03 [2]
##  desc           1.3.0      2021-03-05 [2]
##  devtools       2.3.2      2020-09-18 [2]
##  digest         0.6.27     2020-10-24 [2]
##  dplyr        * 1.0.5      2021-03-05 [2]
##  ellipsis       0.3.1      2020-05-15 [2]
##  entropy        1.2.1      2014-11-14 [1]
##  evaluate       0.14       2019-05-28 [2]
##  fansi          0.4.2      2021-01-15 [2]
##  farver         2.1.0      2021-02-28 [2]
##  fastmap        1.1.0      2021-01-25 [2]
##  fastmatch      1.1-0      2017-01-28 [2]
##  forcats      * 0.5.1      2021-01-27 [2]
##  fs             1.5.0      2020-07-31 [2]
##  FSelector    * 0.33       2021-02-16 [1]
##  gbm            2.1.8      2020-07-15 [2]
##  generics       0.1.0      2020-10-31 [2]
##  GGally       * 2.1.1      2021-03-08 [1]
##  ggfortify    * 0.4.11     2020-10-02 [2]
##  ggplot2      * 3.3.3      2020-12-30 [2]
##  glue           1.4.2      2020-08-27 [2]
##  gridExtra      2.3        2017-09-09 [2]
##  gtable         0.3.0      2019-03-25 [2]
##  haven          2.3.1      2020-06-01 [2]
##  highr          0.8        2019-03-20 [2]
##  hms            1.0.0      2021-01-13 [2]
##  htmltools      0.5.1.1    2021-01-22 [2]
##  httr           1.4.2      2020-07-20 [2]
##  jquerylib      0.1.3      2020-12-17 [2]
##  jsonlite       1.7.2      2020-12-09 [2]
##  knitr          1.31       2021-01-27 [2]
##  labeling       0.4.2      2020-10-20 [2]
##  lattice        0.20-41    2020-04-02 [2]
##  lifecycle      1.0.0      2021-02-15 [2]
##  lubridate      1.7.10     2021-02-26 [2]
##  magrittr       2.0.1      2020-11-17 [2]
##  MASS         * 7.3-53.1   2021-02-12 [2]
##  Matrix         1.3-2      2021-01-06 [2]
##  memoise        2.0.0      2021-01-26 [2]
##  mlbench      * 2.1-3      2021-01-29 [1]
##  mlr          * 2.19.0     2021-02-22 [2]
##  mmpf         * 0.0.5      2018-10-24 [2]
##  modelr         0.1.8      2020-05-19 [2]
##  munsell        0.5.0      2018-06-12 [2]
##  parallelMap    1.5.0      2020-03-26 [2]
##  ParamHelpers * 1.14       2020-03-24 [2]
##  parsnip      * 0.1.5      2021-01-19 [2]
##  parttree     * 0.0.1.9000 2021-03-14 [1]
##  pillar         1.5.1      2021-03-05 [2]
##  pkgbuild       1.2.0      2020-12-15 [2]
##  pkgconfig      2.0.3      2019-09-22 [2]
##  pkgload        1.2.0      2021-02-23 [2]
##  plyr           1.8.6      2020-03-03 [2]
##  prettyunits    1.1.1      2020-01-24 [2]
##  processx       3.4.5      2020-11-30 [2]
##  ps             1.6.0      2021-02-28 [2]
##  purrr        * 0.3.4      2020-04-17 [2]
##  R6             2.5.0      2020-10-28 [2]
##  randomForest   4.6-14     2018-03-25 [2]
##  RColorBrewer   1.1-2      2014-12-07 [2]
##  Rcpp           1.0.6      2021-01-15 [2]
##  readr        * 1.4.0      2020-10-05 [2]
##  readxl         1.3.1      2019-03-13 [1]
##  remotes        2.2.0      2020-07-21 [2]
##  reprex         1.0.0      2021-01-27 [2]
##  reshape        0.8.8      2018-10-23 [1]
##  rJava        * 0.9-13     2020-07-06 [2]
##  rlang          0.4.10     2020-12-30 [2]
##  rmarkdown      2.7        2021-02-19 [2]
##  rpart        * 4.1-15     2019-04-12 [2]
##  rpart.plot   * 3.0.9      2020-09-17 [1]
##  rprojroot      2.0.2      2020-11-15 [2]
##  rstudioapi     0.13       2020-11-12 [2]
##  rvest          1.0.0      2021-03-09 [2]
##  RWeka          0.4-43     2020-08-23 [1]
##  RWekajars      3.9.3-2    2019-10-19 [1]
##  sass           0.3.1      2021-01-24 [2]
##  scagnostics  * 0.2-4.1    2018-04-04 [1]
##  scales         1.1.1      2020-05-11 [2]
##  sessioninfo    1.1.1      2018-11-05 [2]
##  stringi        1.5.3      2020-09-09 [2]
##  stringr      * 1.4.0      2019-02-10 [2]
##  survival       3.2-7      2020-09-28 [2]
##  testthat       3.0.2      2021-02-14 [2]
##  tibble       * 3.1.0      2021-02-25 [2]
##  tidyr        * 1.1.3      2021-03-03 [2]
##  tidyselect     1.1.0      2020-05-11 [2]
##  tidyverse    * 1.3.0      2019-11-21 [2]
##  usethis        2.0.1      2021-02-10 [2]
##  utf8           1.2.1      2021-03-12 [2]
##  vctrs          0.3.6      2020-12-17 [2]
##  withr          2.4.1      2021-01-26 [2]
##  xfun           0.22       2021-03-11 [2]
##  XML            3.99-0.5   2020-07-23 [2]
##  xml2           1.3.2      2020-04-23 [2]
##  yaml           2.2.1      2020-02-01 [2]
##  source                                  
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.1)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  Github (grantmcdermott/parttree@9d25d2c)
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.4)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.2)                          
##  CRAN (R 4.0.0)                          
##  CRAN (R 4.0.0)                          
## 
## [1] /Users/shiringlander/Library/R/4.0/library
## [2] /Library/Frameworks/R.framework/Versions/4.0/Resources/library